"There is practically no uncertainty. Building on top of ChatGPT kills Novelty Uncertainty. I think everyone agrees on that. But the wild early success of so many generative AI products also kills Complexity Uncertainty! If you build a generative AI product that starts to take off, you’re going to be attacked by other startups, bootstrapped companies, solo builders, and incumbents all hungry for your users."
I think this is fair if you're talking about Right Now but not if we're talking about The Future.
In the context of building moats, Right Now is probably all that matters in this case because of your examples of Jasper/Mutiny and others in other industries when they were early in their hype cycles. That said, there's no way to predict what the AI landscape will look like next year, much less in 5 years. You mention a 7-10-year time horizon for a startup exit -- what could happen in that time? It's not out of the realm of possibility that we could have paradigm-shifting advancements in AI land or, alternatively, some version of a Butlerian jihad. To me, the Complexity Uncertainty question isn't just, "Can we make a business out of this idea?" It's really more like, "Can we create a lasting business that can adapt to whatever the future holds?"
Complexity Uncertainty comes from the adaptation and emergence that make predicting the behavior of complex systems impossible (maybe not in a few years with advancements in AI??). In my view, these companies are dealing with massive amounts of complexity, which will inevitably lead to uncertainty & unpredictable outcomes. Can a short-term moat really protect Gen AI startups and the funds investing in them from complexity uncertainty? I don't think so, but who knows.
Would love to know your thoughts (and Jerry Neumann's)!
Probably my favourite piece of yours so far. Great writing. This has given me the hindsight needed to look back at the various startups I’ve worked at over the years and see why some succeeded whilst others failed. The successes seem to be the ones that managed to exit whilst the novelty uncertainty was still in play.
Jul 18, 2023·edited Jul 18, 2023Liked by Packy McCormick
Great timing and summary on a deceptively simple idea. Because start-ups are all-consuming, this is one of those things that companies file under "important but not urgent". There's always tomorrow to build a moat -- until there isn't.
This applies to a lot of start-ups in emerging economies, including Southeast Asia. Many of these companies are working on ideas which have business models proven in the West, and raising money to be first in the land grab is the priority. Moats are an afterthought, until there's no more money left to raise and spend. That's when you realise you need strategy and not just tactics.
Wow--Gen2---Even as simply a commercial, that's overwhelming! Unfortunately, I own stock in a company called Runway, it trades on an exchange under the symbol RWAY. It's a BDC with nothing to do with AI....but it pays 12%... I guess I need more uncertainty.
Great piece on a topic few can describe with much clarity.
Looking forward to the strategy piece. It's definitely defensible. In the best cases, VCs can help play the strategy role when a full time hire doesn't yet make sense.
"There is practically no uncertainty. Building on top of ChatGPT kills Novelty Uncertainty. I think everyone agrees on that. But the wild early success of so many generative AI products also kills Complexity Uncertainty! If you build a generative AI product that starts to take off, you’re going to be attacked by other startups, bootstrapped companies, solo builders, and incumbents all hungry for your users."
I think this is fair if you're talking about Right Now but not if we're talking about The Future.
In the context of building moats, Right Now is probably all that matters in this case because of your examples of Jasper/Mutiny and others in other industries when they were early in their hype cycles. That said, there's no way to predict what the AI landscape will look like next year, much less in 5 years. You mention a 7-10-year time horizon for a startup exit -- what could happen in that time? It's not out of the realm of possibility that we could have paradigm-shifting advancements in AI land or, alternatively, some version of a Butlerian jihad. To me, the Complexity Uncertainty question isn't just, "Can we make a business out of this idea?" It's really more like, "Can we create a lasting business that can adapt to whatever the future holds?"
Complexity Uncertainty comes from the adaptation and emergence that make predicting the behavior of complex systems impossible (maybe not in a few years with advancements in AI??). In my view, these companies are dealing with massive amounts of complexity, which will inevitably lead to uncertainty & unpredictable outcomes. Can a short-term moat really protect Gen AI startups and the funds investing in them from complexity uncertainty? I don't think so, but who knows.
Would love to know your thoughts (and Jerry Neumann's)!
This is the best explanation of moats I have ever read, for sure. That point about the diminishing uncertainty is so important
Probably my favourite piece of yours so far. Great writing. This has given me the hindsight needed to look back at the various startups I’ve worked at over the years and see why some succeeded whilst others failed. The successes seem to be the ones that managed to exit whilst the novelty uncertainty was still in play.
Great timing and summary on a deceptively simple idea. Because start-ups are all-consuming, this is one of those things that companies file under "important but not urgent". There's always tomorrow to build a moat -- until there isn't.
This applies to a lot of start-ups in emerging economies, including Southeast Asia. Many of these companies are working on ideas which have business models proven in the West, and raising money to be first in the land grab is the priority. Moats are an afterthought, until there's no more money left to raise and spend. That's when you realise you need strategy and not just tactics.
This was great, thanks for this. Very useful.
👀
I was at a Solana Hacker House last year and they said “you’re moat is that you’re just better at shipping.“
Standing here with a shovel, wondering where to start digging. Foreman? Help?
Damn, that's a great point: if the idea is great, the Big Guns are going to be coming for you.
Wow--Gen2---Even as simply a commercial, that's overwhelming! Unfortunately, I own stock in a company called Runway, it trades on an exchange under the symbol RWAY. It's a BDC with nothing to do with AI....but it pays 12%... I guess I need more uncertainty.
Great, snappy piece. Giving me a lot to think about while I get my next one out about Kindred.
I have nothing to add to what was written about in this article (for once)
But I do have a question: how do moats figure in an era of increased anti-monopoly activity?
AI is not a moat, right?
Great piece on a topic few can describe with much clarity.
Looking forward to the strategy piece. It's definitely defensible. In the best cases, VCs can help play the strategy role when a full time hire doesn't yet make sense.
Finally someone explained moat in a way I can truly understand!
Great write up, appreciate the effort. Now back to the tent to think.